¹ Korsgaard's reflections on this topic are presented most recently in a paper written for a special APA session on the state of moral philosophy at the turn of a new century; see Christine Korsgaard, ‘Realism and Constructivism in 20^th Century Moral Philosophy,’ forthcoming in the Journal of Philosophical Research. Page numbers in the text to follow refer to passages in Korsgaard's manuscript.

² Cf. Harman, Gilbert, The Nature of Morality (New York: Oxford University Press, 1977)Google Scholar, chap. 1 and Mackie, J. L., Ethics (Harmondsworth: Penguin Books, 1977), chap. 1.Google Scholar

³ See Boyd, Richard, ‘How to be a Moral Realist,’ as reprinted in Darwall, Stephen, Gibbard, Allan, and Railton, Peter, eds., Moral Discourse and Practice (New York: Oxford University Press, 1997), pp. 105–35;Google ScholarBrink, David, Moral Realism and the Foundations of Ethics (Cambridge, England: Cambridge University Press, 1989);Google Scholar and Railton, Peter, ‘Moral Realism,’ as reprinted in Darwall, , Moral Discourse and Practice, pp. 137–63.Google Scholar

⁴ See Nagel, Thomas, The View from Nowhere (New York: Oxford University Press, 1985)Google Scholar, chap. 8; Parfit, Derek, ‘Reasons and Motivation,’ The Aristotelian Society, Supplementary Volume 77 (1997), pp. 99–130;Google Scholar and Scanlon, T. M., What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998)Google Scholar, chap. 1. Actually Scanlon is a constructivist of a sort about moral reasons, and a realist about normative reasons of other kinds; one could also be a realist about normativity, but a sceptic about morality, denying that moral considerations constitute genuine reasons for action. For my purposes in what follows, the more important position is general realism about the normative, rather than the specific view that moral reasons are prior to and independent of the will, since it is the more general position that fundamentally contrasts with Korsgaard's brand of constructivism. I shall understand NMR in what follows in this more general sense.

⁵ See Nagel, Thomas, The Possibility of Altruism (Princeton: Princeton University Press, 1970), p. 23.Google Scholar

⁶ Thus as I mentioned above, Scanlon's NMR is a form of realism about normativity in general (including what he would call morality in the broader sense), within which his constructivism about the ‘morality of right and wrong’ (or morality in the narrower sense) is embedded. I discuss this aspect of his position in my paper ‘Scanlon's Contractualism,’ Ethics 112 (2002), pp. 429–70CrossRefGoogle Scholar, sec. 4.

⁷ Especially significant in this connection is her book The Sources of Normativity (Cambridge, England: Cambridge University Press, 1996)Google ScholarPubMed, and her 2002 John Locke lectures on ‘Self-Constitution: Action, Identity and Integrity.’

⁸ I assume, here and in what follows, that an agent will have reason to do x only if their doing x would be good or valuable in some way.

⁹ The ideas presented in this section are developed more fully in my paper ‘Three Conceptions of Rational Agency,’ Ethical Theory and Moral Practice 2 (1999), pp. 217–42.CrossRefGoogle Scholar

¹⁰ Another way to put this point would be to say that the Humean interpretation of rational motivation seems to imply a kind of noncognitivism about the normative judgments arrived at through practical deliberation. These function not to ascertain independent truths about (say) what we have reason to do, but rather to express our motivating attitudes.

¹¹ The tendency to coherence is appealed to by Smith, Michael, most fully in ‘The Definition of “Moral”,’ in Jamieson, Dale, ed., Singer and his Critics (Oxford: Blackwell, 1999), pp. 38–63;Google Scholar see also his The Moral Problem (Oxford: Blackwell, 1994)Google Scholar, sec. 5.10, and his ‘In Defense of The Moral Problem: A Reply to Brink, Copp, and Sayre-McCord,’ Ethics 108 (1997), pp. 84–119Google Scholar, sees. 5–6. The disposition to do what one believes one ought is discussed by Broome, John, in ‘Reasons and Motivation,’ The Aristotelian Society supplementary volume 77 (1997), pp. 131–46.Google Scholar

¹² These proposals are all due to Velleman, J. David. See his Practical Reflection (Princeton: Princeton University Press, 1989)Google Scholar, for the desire to do what makes sense; his ‘What Happens When Someone Acts?,’ Mind 101 (1992), pp. 461–81Google Scholar, for the desire to act in accordance with reasons; and his ‘The Possibility of Practical Reason,’ Ethics 106 (1996), pp. 694–726Google Scholar, and ‘Deciding How to Decide,’ in Cullity, Garrett and Gaut, Berys, eds., Ethics and Practical Reason (Oxford: Clarendon Press, 1997), pp. 28–52Google Scholar, for the desire for autonomy. The basic strategy of postulating a tendency to rationality that is causally responsible for rational action is anticipated in Hempel, C. G., ‘Rational Action,’ Proceedings and Addresses of the American Philosophical Association 35 (1961–62), pp. 5–24.Google Scholar

¹³ Some might prefer to characterize akrasia differently, as involving action contrary to one's judgment about what it would be best to do. I shall assume, however, that these formulations are equivalent: in the context of practical deliberation, judgments about what it would be good to do are, in effect, judgments about what one has reason to do.

¹⁴ The remarks that follow apply most directly to the versions of metainternalism proposed by Broome and Smith. Velleman's variant is more complicated, since the higher-order desires he postulates are meant to be constitutive not of rational agency in particular, but of agency across the board. (The desire to act for reasons might seem to be a constitutively rational desire that one acts against in cases of akrasia, but only if it is given a de dicto interpretation, which Velleman himself rejects.) Akratic action, on his account, will be motivated by the desire(s) constitutive of agency, and what makes this possible is the fact that the agent's normative beliefs about what it would be best to do fail to engage properly with those constitutive desires. I suspect that this strategy for explaining akrasia ends up tracing it to defects in the agent's normative understanding, in a way that is false to the potential perversity of the phenomenon; but there is not the space to go into this problem here.

¹⁵ Thus the interpretation of ‘can’ in contexts such as the present one is notoriously controversial. Proponents of meta-internalism might draw on the empiricist tradition, contending that akratic agents could have complied with their better judgment just in the sense that they would have so acted had their basal disposition to do what they ought been stronger under the circumstances. Accounts of ‘can’ that take this form suggest a kind of psychological determinism that, I believe, cannot be reconciled with our sense of our own agency. For discussion of this point, see my ‘Moral Responsibility and the Practical Point of View,’ in Beld, Ton van den, ed., Moral Responsibility and Ontology (Dordrecht: Kluwer Academic Publishers, 2000), pp. 25–47.CrossRefGoogle Scholar

¹⁶ Talk of an unreduced ontology of agents and their choices or decisions is likely to raise the puzzling specter of agent-causation, suggesting that we launch our decisions into the world from a position curiously outside of it, as a kind of ‘unmoved mover’ of the things we do. The mistake in this picture, as I see things, is not the failure to reduce agents to congeries of psychological states or attitudes, but the suggestion that the relation between agents and their choices is a causal one. To say that agents cause their decisions or choices is to say that we can explain these attitudes by citing the agent; but the relation between agents and their choices is fundamentally attributional rather than explanatory. For more on this issue, see my ‘Moral Responsibility and the Practical Point of View.’ I would further suggest that this whole conceptual scheme of agents and their choices or decisions has application only to creatures who are capable of practical reason and deliberation, the kind of reflection on normative considerations that is my subject in this part of the present paper.

¹⁷ Compare my Responsibility and the Moral Sentiments (Cambridge, Mass.: Harvard University Press, 1994)Google Scholar, sec. 5.2. Similar ideas are interestingly developed in Korsgaard, ‘The Normativity of Instrumental Reason,’ in Garrett Cullity and Berys Gaut (eds.), Ethics and Practical Reason, pp. 215–54—with the important difference that Korsgaard interprets the stance of volitional commitment in essentially normative terms, whereas I do not. I discuss this difference in my paper Normativity, Commitment, and Instrumental Reason, Philosophers' Imprint 1, no. 3 (December 2001): http://www.philosophersimprint.org/001003.

¹⁸ In discussing this example I sketch the straightforward Humean style of causal explanation rather than the kind to which meta-internalism is committed; but the example could equally be developed in meta-internalist terms. Doing so, we would treat the two agents A and B as motivated by the same formal desire (e.g. to do what they ought), but as differing in their beliefs about what it is they ought to do.

¹⁹ The classic statement of this line of thought is Davidson, Donald, ‘Actions, Reasons, and Causes,’ as reprinted in his Essays on Actions and Events (Oxford: Oxford University Press, 1980), pp. 3–20.Google Scholar

²⁰ Compare Kant's remarks at the beginning of the second section of the Groundwork (p. 407 in the Akademie edition) about the impossibility of knowing with certainty the true motive on which one acts.

²¹ I find myself in agreement here with some points made by Baier, Annette in ‘Actions, Passions, and Reasons,’ as reprinted in her Postures of the Mind (Minneapolis: University of Minnesota Press, 1985), pp. 109–134.Google Scholar I should like to stress, however, that it is no part of volitionalism as I understand it to deny that the intentional states that incorporate our understanding of our reasons can cause the bodily movements we make in acting. What volitionalism denies is not the causal efficacy of such attitudes as choice or decision, but the idea that those attitudes are in turn caused by our antecedent desires and beliefs.

²² I received helpful feedback on earlier versions of this paper from audiences present at the following occasions: an APA session on the future of moral philosophy in December 2001, the 2002 Riverside Conference on the history and future of the will, and the 2002 Royal Institute of Philosophy conference on action and agency. I have particularly benefited from written comments by Michael Bratman, Christine Korsgaard, and Helen Steward.